Copilot & AI Agents for Data Science Bootcamp [2026]

<div>In this hands-on bootcamp, you will master Microsoft CoPilot, GPT-5, and intelligent AI agents for data science. You’ll master the full data science workflow, including data wrangling and feature engineering, data cleaning and merging with CoPilot. We will then cover data visualization and storytelling, turning raw data into dashboards and narratives that drive business decisions. You’ll also cover model development and validation, building and evaluating classifiers while tracking performance using metrics such as accuracy, precision, recall and ROC curves. Finally, you’ll cover anomaly detection, applying methods such as Z-Score and Isolation Forest to spot unusual patterns before they cost money..</div><div><br></div><div>What You’ll Learn:</div><div><ul><li><span style="font-size: 1rem;">Clean and prepare real-world datasets using CoPilot’s advanced prompt engineering.</span></li><li><span style="font-size: 1rem;">Build predictive models for forecasting, classification, and anomaly detection.</span></li><li><span style="font-size: 1rem;">Automate feature engineering and data wrangling tasks with custom AI agents.</span></li><li><span style="font-size: 1rem;">Visualize trends and correlations using Matplotlib, Seaborn, and Plotly inside CoPilot.</span></li><li><span style="font-size: 1rem;">Detect anomalies using Z-Score and Isolation Forest techniques.</span></li><li><span style="font-size: 1rem;">Create executive-level insights and recommendations from raw data.</span></li><li><span style="font-size: 1rem;">Compare and evaluate multiple machine learning models with proper validation.</span></li><li><span style="font-size: 1rem;">Design custom GPTs for advanced analysis, reporting, and business strategy.</span></li></ul></div><div><br></div><div>Bootcamp Modules:</div><div><ul><li><span style="font-size: 1rem;">CoPilot Overview & AI Agents Demo – From messy data cleanup to CEO-level storytelling.</span></li><li><span style="font-size: 1rem;">Data Wrangling & Feature Engineering in CoPilot – Practical workflows for handling missing values, merging datasets, and creating features.</span></li><li><span style="font-size: 1rem;">Data Visualization in CoPilot – Scatter plots, heatmaps, pairplots, and executive-ready dashboards.</span></li><li><span style="font-size: 1rem;">Model Development & Validation – Build, evaluate, and deploy machine learning pipelines.</span></li><li><span style="font-size: 1rem;">Anomaly Detection – Spot unusual trends with Z-Scores and Isolation Forest agents.</span></li></ul></div><div><br></div><div>By the end of this bootcamp, you’ll know how to analyze data and have the skills to build AI-augmented workflows that drive faster, smarter, and more impactful decisions.</div>

What you'll learn:

• Build Data Wrangling AI agents in CoPilot to automate cleaning and preparation tasks on complex datasets.
• Design effective prompts and apply prompting strategies (zero-shot, few-shot, chain-of-thought) to optimize outputs from generative AI systems.
• Use the Pandas library and Microsoft CoPilot to load, manipulate, and analyze real-world datasets programmatically.
• Perform feature engineering tasks such as one-hot encoding, normalization, and standardization to prepare data for machine learning models.
• Apply practical techniques for cleaning messy datasets: handling missing values, removing duplicates, merging data sources, and ensuring consistent formatting.
• Master Data visualization Libraries such as Matplotlib, Seaborn, and Plotly Express to plot static and interactive insight-rich visuals.
• Gain hands-on experience with Microsoft Copilot’s Analyst Agent to automate visualization workflows, generate perspectives quickly, and interpret outputs
• Understand common data visualization types including scatterplots, bubble charts, bar charts, line charts, histograms, box plots, pie charts, and area charts
• Build and interpret regression line plots to study correlations between features and quantify the strength of relationships in data.
• Develop and evaluate classification models (e.g., Logistic Regression, Decision Trees, SVMs, Random Forests, Gradient Boosting, kNN, Naive Bayes)
• Construct and analyze confusion matrices, & calculate key metrics (accuracy, precision, recall, specificity, F1 score, ROC-AUC) to assess model performance
• Identify which performance metrics matter most in specific contexts (e.g., fraud detection vs. marketing campaigns) and justify model selection
• Use CoPilot to build, evaluate, & interpret machine learning pipelines; from exploratory data analysis to model training & evaluation
• Explain the concept of anomaly detection, describe its importance in uncovering unusual patterns, and illustrate real-world applications such as fraud detection
• Apply the Z-score method by calculating and interpreting z-scores, detecting outliers in sales datasets, and visualizing deviations from average performance
• Build an AI Agent in Microsoft Copilot that automates Z-score analysis for sales data, detects anomalies beyond set thresholds, & provides clear visualization
• Implement the Isolation Forest algorithm in Copilot to design an AI Agent (“Isolation Forest Detector”) that isolates and highlights anomalous sales behaviors
• Evaluate the business impact of anomalies uncovered through both techniques, explaining how these insights inform decisions on risks (e.g., revenue drops)